JD Advertising and Huawei Kunpeng report 15% lower inference latency
JD Advertising and Huawei Kunpeng deployed a Kunpeng supernode architecture for generative advertising inference, using UnifiedBus to connect larger memory pools, two-level KV caching and revised RPC communication. The partners reported that the system cut end-to-end inference latency by 15% and handled 5% of traffic in core advertising scenarios during the 618 peak. These performance figures come from the collaborating parties, and the deployment does not yet cover all search, advertising and recommendation workloads.
The report documents a production deployment of generative-AI inference infrastructure under high e-commerce traffic, highlighting memory pooling and data-processing architecture as practical constraints on real-time advertising models.

电商广告的账,最后都记在毫秒上。
京东给全链路的时延预算是300毫秒,留给广告模型推理的部分只有20到100毫秒。超出这个窗口,用户滑动页面时会察觉到卡顿,电商广告成交也就无从谈起。而在最近两年,这套推荐系统正从传统深度学习迁往生成式范式:行为上下文长了10倍,模型结构大了100倍,算力需求随之涨了大约1000倍。机房扩容的周期,却是按季度计的。
当然,这不仅是京东一家所面临的挑战,而是困扰每一家涉及搜广推的互联网企业的共性难题。而在解决这个难点的路上,京东已经先行启程,为行业提供了可借鉴的解题思路。
算力的账:两位数增长与千倍挑战
先看生意本身。京东覆盖超7亿年度活跃用户,平台上70%的商家同时也是广告主,每月帮助顶级品牌触达超4000万核心受众,作为京东集团旗下的营销推广平台,面向B端用户的京准通与C端流量双向协同驱动增长。
#JD Advertising#Huawei Kunpeng#UnifiedBus#KV Cache#Agentic AI